Air-stable organic semiconductors based on 6,6′-dithienylindigo and polymers thereof

Article abstract of DOI:10.1039/c4tc00651h

Herein we report on the synthesis and properties of 6,6′-dithienylindigo (DTI), as well as its solubilized N,N′-di(tert-butoxy carbonyl) derivative (tBOC-DTI). tBOC-DTI can be electropolymerized and thermally interconverted into films of poly(DTI). Thin films of DTI afford quasi-reversible 2-electron reduction and oxidation electrochemistry, and demonstrate ambipolar charge transport in organic field-effect transistors with a hole mobility of up to 0.11 cm² V^-1 s^-1 and an electron mobility of up to 0.08 cm² V^-1 s^-1. Operation of the p-channel shows excellent air stability, with minimal degradation over a 60 day stressing study. Poly(DTI) can be reversibly oxidized and reduced over hundreds of cycles while remaining immobilized on the working electrode surface, and additionally shows a pronounced photoconductivity response in a diode device geometry. This work shows the potential of extended indigo derivatives for organic electronic applications, demonstrating impressive stability under ambient conditions. This journal is

Full text of DOI:10.1039/c4tc00651h

Journal of
Materials Chemistry C
View Article Online
View Journal | View Issue
PAPER
Air-stable organic semiconductors based on
6,60-dithienylindigo and polymers thereof†
Cite this: J. Mater. Chem. C, 2014, 2,
E. D. Głowacki, a D. H. Apaydin,a Z. Bozkurt,a U. Monkowius,b K. Demirak,c E. Tordin,a
8089
*
M. Himmelsbach,d C. Schwarzinger,e M. Burian,f R. T. Lechner,f N. Demitri,g G. Vossa
and N. S. Sariciftcia
Herein we report on the synthesis and properties of 6,60-dithienylindigo (DTI), as well as its solubilized N,N0-
di(tert-butoxy carbonyl) derivative (tBOC-DTI). tBOC-DTI can be electropolymerized and thermally
interconverted into films of poly(DTI). Thin films of DTI afford quasi-reversible 2-electron reduction and
oxidation electrochemistry, and demonstrate ambipolar charge transport in organic field-effect
transistors with a hole mobility of up to 0.11 cm2 Vꢀ1 sꢀ1 and an electron mobility of up to 0.08 cm2 Vꢀ1
s
ꢀ1. Operation of the p-channel shows excellent air stability, with minimal degradation over a 60 day
stressing study. Poly(DTI) can be reversibly oxidized and reduced over hundreds of cycles while
remaining immobilized on the working electrode surface, and additionally shows a pronounced
photoconductivity response in a diode device geometry. This work shows the potential of extended
indigo derivatives for organic electronic applications, demonstrating impressive stability under ambient
conditions.
Received 31st March 2014
Accepted 21st July 2014
DOI: 10.1039/c4tc00651h
www.rsc.org/MaterialsC
minimal performance degradation even aer measuring for >1
year.7 In parallel, there have been extensive studies reported in
Introduction
Indigo and its 6,60-dibromo derivative, Tyrian Purple, have been
used as dyes and pigments for thousands of years by many of
the world civilizations.1 In the 19th century, the high commer-
cial value of indigo dyestuffs encouraged early organic chemists
to explore synthesis routes to indigos, and indeed the economic
race to develop synthetic indigos was a major impetus behind
the development of modern synthetic organic chemistry.2,3
Recently, it was found that indigo,4 6,60-dibromoindigo,5 and
similar hydrogen-bonded molecules6,7 show considerable
promise for organic electronic applications. In eld-effect
transistors, indigos show ambipolar operation with mobility in
the range 0.01–1.5 cm2 Vꢀ1 sꢀ1. These materials stand out in
particular due to excellent operational stability in air, with
the literature on semiconducting properties of the closely
related building blocks of isoindigos8,9 and diketopyrrolo-
pyrroles,10,11 where these chromophores function as electron-
accepting units in donor–acceptor type polymers. Indigos
themselves were overlooked for such applications, as they are
considered to have poor intramolecular p-conjugation.12
Nevertheless, in crystalline thin lms indigos adopt cofacial
packing with sufficient transfer integral between neighboring
molecules to support charge transport.7 Formation of highly
crystalline lms is driven by the interplay of p–p stacking and
intra- and intermolecular hydrogen bonding (H-bonding). Very
recently in this journal, polymers containing indigo as an
acceptor unit in the main chain were reported, with large side
chains attached to nitrogens to preserve solubility. These
materials show promising n-type behavior with an electron
aLinz Institute for Organic Solar Cells (LIOS), Physical Chemistry, Johannes Kepler
University, Linz, Austria. E-mail: eric_daniel.glowacki@jku.at
bInstitute of Inorganic Chemistry, Johannes Kepler University, Linz, Austria
mobility of ꢁ10ꢀ3 cm2 Vꢀ1 sꢀ1
.
The authors discuss the
13
problematic aspect of the solubilizing chains distorting the
planarity of the indigo chromophore. The aim of the study
presented here was to develop an efficient synthesis strategy to
extend the indigo molecule with well-known organic semicon-
ducting moieties – as a simple building block we use here
thiophene – while preserving the H-bonded highly planar
indigo unit. To this end, we present two synthesis routes to
enable solution-based chemistry while yielding the H-bonded
compound in the end. This allows us to test how the indigo
moiety can template crystalline growth when functionalized
with conjugated units. Additionally, substitution with
c
˙
Materials Science and Engineering, Izmir Katip Çelebi University, Izmir, Turkey
dInstitute for Analytical Chemistry, Johannes Kepler University, Linz, Austria
eInstitute for Chemical Technology of Organic Materials, Johannes Kepler University,
Linz, Austria
fInstitute of Physics, Montanuniversitaet Leoben, Franz-Josef-Strasse 18, 8700 Leoben,
Austria
gElettra – Sincrotrone Trieste, S.S. 14 Km 163.5 in Area Science Park, 34149 Basovizza,
Trieste, Italy
† Electronic supplementary information (ESI) available: 5aVat details, cyclic FET
characteristics, MALDI spectra, X-ray structural data, and computational details.
CCDC 1006293. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c4tc00651h
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. C, 2014, 2, 8089–8097 | 8089

View Article Online
Journal of Materials Chemistry C
Paper
thiophenes allows facile polymerization. Insolubility of indigos
Diffraction-quality crystals of 5a were grown in a tempera-
is caused by the interplay of H-bonding and p–p stacking, ture-gradient tube furnace under a stream of dry nitrogen with a
leading to exceptionally high crystal lattice energies. As a tech- ow of 2 L minꢀ1 using the source material that had previously
nique to make indigos tractable for synthesis, N,N0-dis- been puried by sublimation. The source temperature was held
ubstitution with alkyl or acyl units is used to eliminate at 305 ꢂC and crystals formed in a downstream zone with 110 ꢃ
H-bonding and distort the planarity of the original indigo 15 ꢂC. Crystals were measured using synchrotron radiation with
˚
molecule. Though such substitution facilitates chemical a wavelength of 0.7 A. Experimental details can be found in the
manipulation, it also interrupts the crystalline packing inter- ESI.† Projections of the crystal packing structure viewed down
actions characteristic of H-bonded indigos and therefore is the a and b axes are shown in Fig. 1. The indigo core of 5a is, like
likely detrimental to transport. In this work we introduce ther- other H-bonded indigos, highly planar, while the thiophene
molabile protection groups: t-butoxy carbonyl (tBOC) units. rings are at a torsion angle of 14ꢂ relative to the indigo plane.
This can render the indigo molecule soluble for chemical Intermolecular H-bonding occurs along the a-direction, with a
manipulation and processing.14 Aerwards, the tBOC groups NH/O ¼ distance of 2.51 A (Fig. 1b). Typically, this distance for
˚
0
˚
can be decomposed into gaseous products CO2 and isobutene indigo and 6,6 -disubstituted indigos is shorter: ꢁ2.1 A. Along
by simple heating to temperatures >150ꢂ thereby regenerating the b-direction, there is pronounced p–p stacking, with an
˚
the planar, H-bonded form of indigo. Using this transient intermolecular centroid–centroid distance of 4.78 A (Fig. 1a).
protection technique, we can synthesize 6,60-dithienylindigo The packing arrangement of 5a is quite unlike any reported
(DTI, 5a) which can be used to fabricate ambipolar OFETs with indigo crystal structures. Typically, indigos crystallize in a criss-
mh ¼ 0.11 and me ¼ 0.08 cm2 Vꢀ1 sꢀ1. These OFETs demonstrate cross arrangement where each molecule forms single hydrogen
impressive stability in air during a 60 day stressing study. The bonds to four neighbouring molecules, with p–p stacking
electropolymerized form poly(DTI) is an insulator, however it occurring in the perpendicular direction. The one exception is
shows a strong (resistance drop ꢁ 104) photoconductivity 6,60-dichloroindigo, which forms double hydrogen-bonds to two
response in diode geometry. The poly(DTI) can also undergo neighbours, leading to linear chains of molecules with adjacent
fully reversible reduction at low potentials (ꢀ480 mV vs. Ag/ chains propagating in the same direction and a brick-wall type
AgCl) over hundreds of repeated cycles, making it the rst p-stacking structure.16 The linear-chain vs. criss-cross H-bond
example to our knowledge of a polymer vat dye that preserves its lattice arrangements are well-known from the work on other H-
characteristic electrochemistry whilst not dissolving in the bonded pigments, where several materials can form both linear-
electrolyte. The two synthesis routes presented here to achieve chain and criss-cross polymorphs under certain conditions.17
extended indigos via coupling reactions open up new avenues in Though the linear-chain motif is present in 5a, the H-bonded
using indigo as a functional and stable building block for sheets of molecules are tilted with respect to each other by 89ꢂ
organic electronic materials.
along the c-axis. The formation of such staggered linear chains
is familiar from p-substituted diketopyrrolopyrrole pigments,
but has not been observed in indigos before.18 The close p–p
stacking along the b-axis probably provides the primary
contribution to electronic transport through lms of 5a.
Results and discussion
Synthesis and crystal structure
Initially, we attempted coupling reactions on 2 directly. However,
due to extremely poor solubility of the compound, we were not
able to obtain coupling products. Two synthesis approaches that
Optical properties
overcome the solubility problem, shown in Scheme 1, were The strong and surprisingly bathochromic absorption in indi-
evaluated in order to synthesize compounds 5a,b. In the route goid dyes originates from the so-called indigoid H-chromo-
shown in the upper part of the scheme, the protect–deprotect phore, a cross-conjugated arrangement of electron donor and
route, we synthesize the tBOC protected form 4 of Tyrian Purple acceptor groups.19 Thus, low-energy visible light absorption
2. Compound 4 is transformed to tBOC-DTI 5b via Stille or Suzuki occurs due to charge transfer from the electron-rich NH groups
coupling with two equivalents of 2-(tributylstannyl)thiophene or to the electron-poor carbonyl functions. In order to better
2-thienyl boronic acid, respectively. In the alternate route, Stille understand the absorption and emission data shown in Fig. 2,
coupling is performed directly on 4-bromo-2-nitrobenzaldehyde we conducted DFT calculations to estimate molecular confor-
1, generating 3 with essentially quantitative yield. Compound 3 is mation and frontier orbital energies (Fig. 3). The characteristic
then transformed via the Baeyer–Drewsen reaction15 into DTI 5a frontier orbital shapes of the cross-conjugated H-chromophore
with a yield of approx. 25%, which can in turn be protected with are the same in compounds 4 and 5 as previously calculated by
tBOC to yield 5b. Material 5a could be isolated directly from the DFT for other indigos.20,21 UV-Vis absorption spectra of 5a and
dried reaction mixture by temperature gradient sublimation in a 5b in DMSO solution are shown in Fig. 2a. Though 5b is highly
vacuum to yield highly pure samples. Compound 5b, on the other soluble in a wide range of organic solvents, DMSO was chosen
hand, was puried by column chromatography to separate it as a common solvent that also allowed, aer heating, dissolu-
from small amounts of 4 and mono-thienyl coupling byproduct tion of the deprotected form 5a. Compound 5a shows a main
5c. Stable aqueous vats of the reduced form 5avat could be easily HOMO–LUMO transition peak at 580 nm, with a lower-energy
prepared using sodium dithionite as the reducing agent despite shoulder at ꢁ625 nm, causing the solution to appear deep
the hydrophobic thiophene substituents.
purple in colour. By way of comparison, indigo in DMSO has a
8090 | J. Mater. Chem. C, 2014, 2, 8089–8097
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry C
Scheme 1 Two synthesis routes to obtain 6,60-dithienylindigo, and its polymerized form via electrochemical polymerization. (i) acetone, 2M
KOH, ꢀ5 ꢂC to rt. (ii) Pd(PPh3)4, 2-(tributylstannyl)thiophene, toluene, reflux under N2. (iii) Boc2O, DMAP, THF, rt under N2. (iv) 2-thienylboronic
acid, Pd(dppf)Cl2, 1M K2CO3, THF, reflux under N2. (v) Na2S2O4, 2M KOH, heat.
Fig. 2 (a) UV-Vis absorption of 5a,b in DMSO solution. Cuvettes show
0.1 mM solutions. (b) UV-Vis absorption of thin films on ITO glass. (c)
Fig. 1 Crystal structure of DTI (5a). (a) A slightly tilted view down the b-
Photoluminescence and excitation spectra of 5b in DMSO solution. (d)
axis, showing p–p stacking interactions between H-bonded chains of
Photoluminescence and excitation spectra of 5a in DMSO.
indigos. (b) View down the a-axis, showing the direction of propaga-
tion of the H-bonded linear chains.
absorption, with a peak at 540 nm, due to the added contribu-
tion of the electron-withdrawing tBOC groups, which decrease
lmax at 620 nm. It is known that electron-rich substituents in the the electron-donating ability of the nitrogen atoms within the p-
6,60 positions inductively push electron density onto the system of the H-chromophore. Compound 5b shows photo-
carbonyl groups,22–24 decreasing their electron-accepting char- luminescence with two distinct peaks, at 500 and 610 nm. The
acter and therefore hypsochromically shiing the absorption of former originates from the chromophore containing the thio-
the indigoid H-chromophore. It is not surprising that the highly phene-phenyl system, as can be deduced from the excitation
electron-rich thiophene functions shi the absorption so spectrum. The lower energy luminescence at 610 nm can be
strongly to the blue in the case of 5a. The protected form of the attributed to the H-chromophore. The excitation prole for this
compound, 5b, has an even more shied H-chromophore luminescence band follows that measured in the absorbance
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. C, 2014, 2, 8089–8097 | 8091

View Article Online
Journal of Materials Chemistry C
Paper
Fig. 3 Optimised ground-state geometries and frontier orbital energies of compounds 4, and 5a,b and cut-out segments of polymers 6a and 6b
obtained by DFT calculations.
spectrum, however with a more distinct ne structure, clearly electrochemical oxidation and reduction onsets, respectively
showing a splitting between higher and lower energy bands that (Table 1).
is visible in the case of the unprotected 5a. The luminescence
In the case of thin lms, the situation follows what is
from 5a, in solution however, shows one single peak at 676 nm, observed in solutions (Fig. 2b). In evaporated lms of 5a, the
with the excitation prole tting closely to the absorbance absorption splits into a lower energy band with a peak at 670 nm
spectrum. These differences with respect to the protected and a higher energy band at 552 nm, the latter having a higher
compound can be rationalized by the fact that tBOC substituted oscillator strength. Thus, the lms appear violet in colour and
compounds are known to be highly distorted with respect to the are identical to 6,60-dibromoindigo (2) in appearance. Electro-
central double bond,14 thus the indole–thiophene chromophore polymerized lms with protecting groups (6b) are yellow in
can be electronically decoupled from the central H-chromo- colour, with only one absorption peak at 398 nm, indicating that
phore. In the rigid and planar 5a, the electronic structure the indigoid moieties are highly twisted and thus have limited
collapses to lower energy H-chromophore states. Qualitatively conjugation. This is also reected in the DFT conformation
consistent with these experimental results are the results of the calculation. Upon heating and deprotection, the deprotected
DFT calculations of frontier orbital energies: the presence of the form of the polymer (6a) regains a low energy absorption broad
electron-poor tBOC group energetically stabilizes both the peak at 632 nm, and appears green in colour. Neither polymer
HOMO and LUMO orbitals and leads to an increase of the lm showed any detectable photoluminescence.
energy gap compared to the unprotected compounds. Fig. 3
shows the relative predicted energy levels as well as ground-
Electrochemistry and electropolymerization
state optimized conformations. The predicted DFT band gaps
are systematically overestimated, however the qualitative
differences, such as the decrease of band gap, upon depro-
tection (5b / 5a, 6b / 6a) reect what is found experimentally.
Also the relative positions of the calculated HOMO and LUMO
energy levels follow the same trends found from
Cyclic voltammetry was used to evaluate the electrochemical
behaviour of 5a,b and 6a,b (Fig. 4). In all cases, 0.1 M tetrabu-
tylammonium hexauorophosphate in acetonitrile was used as
the electrolyte solution, with Ag/AgCl functioning as the quasi-
reference electrode, and a Pt foil as the counter electrode. Using
Table 1 Estimations of frontier orbital energy levels
HOMOa (eV)
LUMOb (eV)
Eg(CV) (eV)
Eg
(eV)
(opt)
Material
Eox (V vs. Ag/AgCl)
Ered (V vs. Ag/AgCl)
2
1.10
0.96
1.30
1.11
1.25
ꢀ0.86
ꢀ0.78
ꢀ0.39
ꢀ0.48
ꢀ0.60
ꢀ5.8
ꢀ5.6
ꢀ6.0
ꢀ5.8
ꢀ5.9
ꢀ3.8
ꢀ3.9
ꢀ4.3
ꢀ4.2
ꢀ4.0
2.0
1.7
1.7
1.6
1.9
2.0
1.8
2.0
1.7
2.3
5a
5b
6a
6b
a
b
EHOMO ¼ ꢀ(E[onset,ox. vs. NHE] + 4.75) (eV). ELUMO ¼ ꢀ(E[onset,red. vs. NHE] + 4.75) (eV).
8092 | J. Mater. Chem. C, 2014, 2, 8089–8097
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry C
ferrocene as a standard,25 we estimated frontier orbital energies dry N2 environment. Both potentiodynamic and potentiostatic
for all compounds, summarized in Table 1 together with band oxidative polymerization were conducted. Electropolymerization
gaps calculated from electrochemistry and optical absorption. A of the monomer 5b was realized potentiodynamically by cycling
worthwhile observation is that the optical band gaps are always potentials between 1600 mV and ꢀ500 mV with a scan rate of
equal to or greater than the electrochemically calculated band 50 mV sꢀ1 (Fig. 4). Throughout the polymerization a shi in the
gaps. In organic semiconducting materials typically this rela- oxidation potential to higher values was observed which we
tion is reversed. The exact origin of the unusually small elec- attribute to the increased ohmic loss as the polymer lm grows.
trochemical gap in indigos, observed in the compounds tBOC-protected poly(DTI) 6b could then be deprotected by
reported here as well as previously for unsubstituted indigo,4 is simply heating at 150 C for 5 minutes, yielding lms of poly-
ꢂ
unclear. It may be related to the relatively polar indigo system (DTI) 6a. Fig. 5c and d show AFM micrographs of poly(DTI) lms
which can easily undergo oxidation and reduction processes.
before and aer deprotection. In both cases a crystalline grain-
We exploited the pendant thiophene groups on DTI 5b to like morphology is visible, quite similar to lms of monomer 5a
conduct oxidative polymerization to yield 6b immobilized on an (Fig. 5a and b) though with signicantly lower roughness.
ITO working electrode. DTI 5b was dissolved to a concentration Potentiostatic polymerization yielded lms with identical
of 0.05 M in an electrolyte solution of 0.1 M tetrabutylammo- morphologies, however signicantly thicker (>1 mm) lms could
nium hexauorophosphate in acetonitrile. An Ag/AgCl wire be grown since the material was not de-doped by cycling and
functioned as the quasi-reference electrode and a Pt foil served could be conductive enough to support the growth of thicker
as the counter electrode. The experiments were carried out in a lms. Neither 6a nor 6b could be solubilized for analysis. The
structure was conrmed using matrix-assisted laser-desorption
ionization (MALDI) mass spectroscopy using samples of thick
potentiostatically grown lms which could be scratched off the
electrode surface as akes (see ESI† for spectra).
Indigoids are vat dyes, thus upon reduction or oxidation 5a
becomes highly soluble in polar solvents. This causes rapid
dissolution of lms during cyclic voltammetry and impedes the
measurement of respective re-oxidation or re-reduction. In
contrast to lms of monomer 5a, the polymerized 6a showed
extraordinary cyclic stability. We conducted 500 scans in
acetonitrile under ambient conditions, and found no degrada-
tion of the lm or change in electrochemical behaviour, the
successive scans could be overlaid. Oen, repeated reductive
electrochemistry in air leads to degradation, however samples
of 6a were highly stable. We attempted measurement in
aqueous electrolytes and found that no electrochemical current
Fig. 4 Comparison of the cyclic voltamograms of 5b measured in
solution, 5a measured as a film, and polymers 6a and 6b as films. A Fig. 5 AFM micrographs of (a) 80 nm of 5a on ITO, (b) 80 nm of 5a on
scan rate of 20 mV sꢀ1 was used in all cases. The bottom graph shows AlOx/TTC, (c) a 140 nm film of 6b polymerized on ITO, and (d) the same
the potentiodynamic electropolymerization procedure for 5b / 6b.
film after heating/deprotection, yielding 6a.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. C, 2014, 2, 8089–8097 | 8093

View Article Online
Journal of Materials Chemistry C
Paper
was measurable, which was attributed to the extreme hydro-
phobicity of the 6a lm. This has been reported for other
conjugated polymers as well. Replacing the electrolyte with the
original acetonitrile one immediately yielded the previous
electrochemical behaviour. Similarly, monomer 5a could not be
measured with aqueous electrolytes. It was possible, however, to
make a basic vat of 5a using sodiꢂum dithionite as the reducing
agent. Heating 5a powder to 85 C in 2 M KOH with Na2S2O4
resulted in a dark orange-coloured vat of the reduced form of 5a
(see spectra in the ESI†). If oxygen was excluded, the vat was
stable in its reduced form at least for several days. Admission of
air resulted in reoxidation to the neutral 5a which reprecipitated
out of the solution.
Transistors and diodes
DTI (5a) could be sublimed in a vacuum to form device-quality
lms. AFM micrographs of 80 nm of 5a evaporated with a rate of
0.2 A sꢀ1 on ITO glass and AlOx/TTC are shown in Fig. 5a and b.
˚
Crystalline grains with roughly 100 nm domain size form on
ITO, while on the low-surface energy TTC crystallites are
considerably larger, with domains of several hundred nano-
metres. Devices fabricated using 5b followed by heating to yield
5a failed to give lms, instead isolated crystallites were formed.
This behaviour we previously found with other latent indigo
pigments.14 We fabricated organic eld-effect transistors using
DTI in a device conguration that is known to support transport
in H-bonded pigment-forming molecules. A gate electrode of
evaporated aluminium is used, with a 32 nm layer of anodized
alumina passivated with a 20 nm layer of tetratetracontane
Fig. 6 DTI (5a) OFET transfer characteristics (a) with the inset (b)
showing the device architecture used for the measurement. (c) OFET
device lifetime study – mobility as a function of measurement days in
air. Each day 150 cycles were applied.
(C44H90, TTC) functioning as the composite gate dielectric indicating that O2 causes electron trapping rather than irre-
(Fig. 6b).26 We have reported extensively properties of this gate versible degradation of the material. The 150 cycle procedure
structure in the past.7 Next, 80 nm of DTI are evaporated on top was repeated over 60 days. Between measurements the samples
with a rate of 0.1 A sꢀ1. Finally, top Au source/drain contacts are were stored in an ambient laboratory environment, including
˚
evaporated through a shadow mask to give a W/L of 2 mm/60 room lighting. Over this 60 day stressing study, the hole
mm. Device characteristics were measured in a dry N2 environ- mobility increases slightly over the rst few days, and then
ment using different VDS values. In all cases the FETs showed gradually decays to 0.07 cm2 Vꢀ1 sꢀ1 (Fig. 6c).
ambipolarity (Fig. 6a). A p-type operation is easily obtained with
the saturation hole current being independent of VDS
Diodes using evaporated DTI as well as the poly(DTI) 6a,b
were fabricated using ITO as a transparent substrate electrode.
,
evidencing a low contact resistance for hole injection. From the In the former case, a 140 nm layer of DTI was evaporated, fol-
saturation regime, we calculate mh ¼ 0.1 cm2 Vꢀ1 sꢀ1. The n- lowed by an aluminium top electrode. For poly(DTI) diodes, a
channel, however, was strongly dependent on VDS, therefore a 140 nm lm was prepared by potentiodynamic electro-
substantial contact resistance existed for electron injection. polymerization of 5b, followed by extensive washing with
Once a sufficiently high VDS of 16 V was applied, the n-channel acetonitrile and water to remove residual salts, and nally
saturates, allowing an estimation of me ¼ 0.08 cm2 Vꢀ1 sꢀ1
.
heating to generate a lm of 6a. An evaporated top electrode of
Application of different contact metals aluminium and silver Al or MoOx/Ag completes the device. Current–voltage charac-
lowered the contact resistance for electrons, though with a teristics of DTI 5a and poly(DTI) 6a diodes are shown in Fig. 7.
detriment for hole injection. Based on previous observations Films of 6b were completely insulating in all cases and thus
that indigo and other H-bonded pigments operate stably in air,6 results are not shown. DTI demonstrated essentially no recti-
we conducted stressing/lifetime experiments for DTI FETs, cation, and relatively low currents. Based on the conductivity of
using a pool of eight identical devices. Cyclic measurements of these sandwich geometry devices, the carrier mobility of the
150 scans were applied in air, with a scan rate of 300 mV sꢀ1
.
majority carrier cannot exceed the 10ꢀ3 cm2 Vꢀ1 sꢀ1 range.
Over 150 cycles the device p-channel shows no fatigue in terms Irradiation with light from a solar simulator (80 mW cmꢀ2
)
of mobility, on/off ratio, or hysteresis (see ESI†). The n-channel, results in a pronounced photovoltaic effect and also a roughly
visible on the rst scan, rapidly decays, which is not surprising tenfold increase in conductivity. Considering the markedly
due to the high energetic position of the LUMO. The n-channel higher mobility in FET devices, it is likely that a high trap
recovers when measured in
a N2 atmosphere, however, density exists in this material. This is supported by the
8094 | J. Mater. Chem. C, 2014, 2, 8089–8097
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry C
prepared by intensive mixing of a small amount of the poly-
merized lm together with an excess of dithranol in a mortar. A
small aliquot of the solid mixture was then transferred onto a
MALDI target and rubbed in with a spatula. An average of at
least 500 spectra was summed up to minimize errors due to the
rather rough surface of the solid sample preparation.
Electrochemical measurements were carried out in a dry N2
environment using dry 0.1 M tetrabutylammonium hexa-
uorophosphate in acetonitrile as an electrolyte. Thin lms of
the material on ITO functioned as the working electrode, plat-
inum foil was used as the counter electrode, and Ag/AgCl was
used as the quasi-reference electrode. Measurements were
calibrated against ferrocene. The values of Fc/Fc+ vs. NHE and
NHE vs. vacuum level used in this work were 0.64 and ꢀ4.75 V,
respectively.25 ITO glass slides used for electrochemistry and
diode fabrication were cleaned sequentially with acetone, iso-
propanol, detergent, and DI water, and nally treated with O2
plasma. Aluminum/aluminum oxide/TTC gate structures were
prepared according to reported methods.30,31
Fig.
measured in the dark or under simulated solar illumination with an
intensity of 80 mW cmꢀ2
7 J–V characteristics of ITO/DTI or poly(DTI)/Al devices,
.
photoconductivity effect explained by photogenerated carriers
lling trap states. These observations are consistent with the
similar disparity between diodes and FET devices reported in
other hydrogen-bonded pigment semiconductors, suggesting
an important role of traps in this materials class.27 In the case of
poly(DTI) 6a devices, the effect of photoconductivity is more
dramatic. The devices in the dark have a resistivity of ꢁ10 TU
cm. Irradiation causes an increase in conductivity by 104. The
high resistivity of the poly(DTI) material is explained by the lack
of conjugation along the polymer chain, i.e. the indigo units can
be seen as interrupting along-chain conduction. The excited
state of indigo, however, is known to exist in the enol tautomeric
form, which indeed supports a mesomeric structure allowing
along-chain conjugation (similar to mesomeric structure of
5avat). This latent conjugation phenomenon model was
proposed for indigo-containing polymers by Yamamoto.28 We
hypothesize that this could be the reason for the observed
photoconductivity.
2-Nitro-4-(2-thienyl)benzaldehyde (3). To
a solution of
458 mg (1.9 mmol) of 1 in 60 ml of degassed toluene, 0.6 ml
(2.0 mmol) of 2-(tributylstannyl)thiophene was added. Aer 15
min, 11 mg of tetrakis-triphenylphosphine-palladium(0) was
added. The mixture was stirred for 16 h at 110 ꢂC. Evaporation
of the solvent gave 500 mg of crude 3; for analysis a sample was
sublimed using a source temperature of 85 ꢂC at a vacuum of 1
1
ꢄ 10ꢀ6 mbar, to give a bright yellow powder. H-NMR (CDCl3):
7.20 (dd, 3J ¼ 5.0 Hz, 3J ¼ 3.7 Hz, 1H, 4-Hthienyl), 7.51 (dd, 3J ¼ 5.0
Hz, 4J ¼ 1.0 Hz, 1H, 5-Hthienyl), 7.57 (dd, 3J ¼ 3.7 Hz, 4J ¼ 1.0 Hz,
1H, 3-Hthienyl), 7.98 (dd, 3J ¼ 8.1 Hz, 4J ¼ 1.0 Hz, 1H, 5-H), 8.03
(d, 3J ¼ 8.1 Hz, 1H, 6-H), 8.30 (d, 4J ¼ 1.0 Hz, 1H, 3-H), 10.42 (s,
1H, CHO). 13C-NMR (CDCl3): 120.9, 126.4, 128.4, 128.7, 128.9,
129.9, 130.5, 140.2, 140.3, 150.5, 187.3. HR ESI-MS: m/z found
for [M + Na]+ 256.0038, m/z calculated for [M + Na]+ 256.0039.
6,60-Di(2-thienyl)-indigo (5a). 3 (ꢁ1.8 mmol) was dissolved in
65 ml of acetone. Aer cooling to ꢀ5 ꢂC, 120 ml of 2 M KOH was
added drop-wise over 2 h. Aer 16 h stirring at room tempera-
ture, the mixture was ltered. The solid was washed wꢀit6h water
and ethanol and nally sublimed at 245 ꢂC/1 ꢄ 10 mbar,
yielding 77 mg (19%) of 5a as a purplish powder. IR: among
others 3107 (brd, NH), 3105, 3064, 3055 (all vw), 1604, 1575,
Experimental
Materials and methods
1
3
4
4-Bromo-2-nitrobenzaldehyde (1) was synthesized from 1,4- 1436 (all s). H-NMR (DMSO-d6): 7.51 (dd, J ¼ 5.0 Hz, J ¼ 1.0
dibromo-2-nitrobenzene (Aldrich) using the regioselective lith- Hz, 2H, 3,30-Hthienyl), 7.30 (d, 3J ¼ 8.1 Hz, 2H, 5,50-H), 7.65–7.68
3
iation/formylation reaction reported by Voss et al.29 Compounds (6H, 4,40,7,70-H and 4,40-Hthienyl), 7.71 (d, J ¼ 4.9 Hz, 2H, 5,50-
2 (ref. 29) and 4 (ref. 14) were obtained according to published
procedures.
Hthienyl), 10.54 (brd s, 2H, NH). HR ESI-MS: m/z found for [M +
1,10-bis(diphenylphosphino)ferrocene]-dichlor- H]+ 427.0565, m/z calculated for [M + H]+ 427.0569.
opalladium(II)
(Pd(dppf)Cl2), tetrakis-triphenylphosphine-
A sample of 5a was transformed with Na2S2O4/NaOD in D2O
palladium(0), 2-tributylstannyl-thiophene and 2-thienylboronic into leuco N,N0-dideutero-6,60-di(2-thienyl)-indigo (5avat) for NMR
acid were commercially available and used as-received analysis according to the method reported by Voss.32
(Aldrich). All solvents were distilled before use. Column chro-
1H-NMR (D2O): 7.08 (m, 2H, 4,40-Hthienyl), 7.32–7.40 (4H, 5,50-
3
matography: Merck silica gel 60 (SiO2; 0.040–0.063 mm). IR [ꢀn in H, 3,30-Hthienyl), 7.36 (m, 2H, 5,50-Hthienyl), 7.49 (d, J ¼ 8.1 Hz,
cmꢀ1]: Shimadzu IRTracer-100. UV-Vis: Perkin-Elmer Lambda 2H, 4,40-H), 7.57 (s, 2H, 7,70-H).
950 with a PMT detector. NMR Bruker 300 (1H: 300 MHz; 13C:
1,10-Bis(tert-butoxycarbonyl)-[6,60-di(2-thienyl)-indolidene
75.5 MHz). For 13C experiments: C13APT (Attached Proton Test indolyden]-3,30-dione [(bis-tBOC-6,60-di(2-thienyl)-indigo; 5b],
using the Jmod pulse program) was used to assist in peak and the mono-substituted byproduct 1,10-bis(tert-butyox-
assignment. MALDI spectra were obtained using a Bruker ycarbonyl)-[6-(2-thienyl)-60-bromo]-indolidene indolyden]-3,30-
Autoex III Smartbeam using linear mode. Samples were dione [bis-tBOC-[6-(2-thienyl)-60-bromo]-indigo; 5c]. Procedure
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. C, 2014, 2, 8089–8097 | 8095

View Article Online
Journal of Materials Chemistry C
Paper
1 – Suzuki coupling:33 To a degassed solution of 440 mg (0.71 interesting photoconductivity behaviour and suggests that the
mmol) of 4 in 50 ml of dimethoxyethane, 227 mg (1.8 mmol) of enolic structure might lead to conjugation. The poly(DTI) shows
2-thienyl-boronic acid was added. Aer 10 min stirring under excellent stability during electrochemical cycling, with a notably
N2, 80 mg (0.1 mmol) of Pd(dppf)Cl2 was added. Then the low reduction potential. The results presented here show the
mixture was heated for 30 min to slight reux (Tbath 90 ꢂC) and 2 promise that extended indigos and indigo-based polymers can
ml of 1 M K2CO3 was added. Stirring and heating were have for organic electronics.
continued for 1 h, and then the mixture was set aside at room
temperature for 16 h. Diethylether (200 ml) was added and the
organic layer was washed with water. Aer evaporation of the
Acknowledgements
solvent, the residue (570 mg, red oil) was chromatographed on We are grateful for support from the Austrian Science Foun-
silica gel with cyclohexane/AcOEt (9 : 1). Starting materials 4 dation, FWF, within the Wittgenstein Prize of N. S. Saricici
(120 mg, 44%, Rf ¼ 0.48), 5b (130 mg, 29%, Rf ¼ 0.21) and 5c Solare Energie Umwandlung Z222-N19 and the Translational
(110 mg, 23%, Rf ¼ 0.32) were isolated as red solids. Recrys- Research Project TRP 294-N19 “Indigo: From ancient dye to
tallization from cyclohexane gave the analytically pure products. modern high-performance organic electronic circuits”.
Procedure 2 – Stille coupling:34 496 mg (0.8 mmol) of 4 was
dissolved in 60 ml of dry toluene. The mixture was degassed by
15 min N2 bubbling. Next, 0.6 ml (2 mmol) of 2-(tributylstannyl)
Notes and references
thiophene was added and N2 bubbling was continued for 15
min. Aer addition of 9.2 mg (0.008 mmol) of tetrakis-triphe-
nylphosphine palladium(0), the mixture was stirred for 16 h at
90 ꢂC. Aer evaporation of volatile products, the residue was
chromatographed on silica gel (Cy/AcOEt, 9 : 1).
1 M. Seefelder, Indigo in culture, science, and technology,
ecomed, Landsberg, Germany, 2nd edn, 1994.
2 H. Schmidt, Chem. Unserer Zeit, 1997, 31, 121–128.
3 E. Steingruber, Indigo and Indigo Colorants in Ullmann's
Encycl. Ind. Chem., 2007.
Starting materials 4 (92 mg, 18%), 5b (170 mg, 33%), 5c
(120 mg, 23%)
4 M. Irimia-Vladu, E. D. Głowacki, P. A. Troshin,
G. Schwabegger, L. Leonat, D. K. Susarova, O. Krystal,
M. Ullah, Y. Kanbur, M. A. Bodea, V. F. Razumov, H. Sitter,
S. Bauer and N. S. Saricici, Adv. Mater., 2012, 24, 375–380.
5 E. D. Głowacki, L. Leonat, G. Voss, M.-A. Bodea, Z. Bozkurt,
A. M. Ramil, M. Irimia-Vladu, S. Bauer and N. S. Saricici,
AIP Adv., 2011, 1, 042132–042137.
6 E. D. Głowacki, M. Irimia-Vladu, M. Kaltenbrunner,
J. G˛asiorowski, M. S. White, U. Monkowius, G. Romanazzi,
G. P. Suranna, P. Mastrorilli, T. Sekitani, S. Bauer,
T. Someya, L. Torsi and N. S. Saricici, Adv. Mater., 2013,
25, 1563–1569.
7 E. D. Głowacki, G. Voss and N. S. Saricici, Adv. Mater., 2013,
25, 6783–6800.
8 E. Wang, W. Mammo and M. R. Andersson, Adv. Mater.,
2014, 26, 1801–1826.
9 R. Stalder, J. Mei, K. R. Graham, L. A. Estrada and
J. R. Reynolds, Chem. Mater., 2014, 26, 664–678.
5b. IR: among others 3107, 3074 (all vw), 2974, 2927 (all w),
1703, 1600, 1435, 1147 (all s). – 1H-NMR (CDCl3): 1.70 (s, tBOC
3
3
CH3, 18H), 7.17 (dd, J ¼ 5.0 Hz, J ¼ 3.6 Hz, 2H, 4,40-Hthienyl),
7.44 (dd, 3J ¼ 5.0 Hz, 4J ¼ 0.9 Hz, 2H, 5,50-Hthienyl), 7.49 (dd, 3J ¼
8.0 Hz, 4J ¼ 1.4 Hz, 2H, 5.50-H), 7.52 (dd, 3J ¼ 3.6 Hz, 4J ¼ 0.9 Hz,
2H, 3,30-Hthienyl), 7.78 (d, 3J ¼ 8.0 Hz, 2H, 4,40-H), 8.33 (d, 4J ¼ 1.4
Hz, 2H, 7,70-H). 13C-NMR (CDCl3): 28.2 (tBOC CH3), 113.5 (7,70-
C), 121.6, 124.7, 125.4, 127.2, 128.6 (4,40- and 5,50-C, and 3,30-,
4,40- and 5,50-Cthienyl) – only the signals of those carbon atoms
attached to hydrogens are given. HR ESI-MS: m/z found for
[M + H]+ 627.1611, m/z calculated for [M + H]+ 627.1618.
5c. (Mono-thienyl byproduct). IR: among others 3120 (vw),
2972, 2924 (all w), 1734, 1672, 1600, 1421, 1141 (all s). 1H-NMR
(CDCl3): 1.64 (s, tBOC CH3, 9H), 1.69 (s, CH3, 9H), 7.16 (dd, 3J ¼
5.1 Hz, 3J ¼ 3.7 Hz, 1H, 4-Hthienyl), 7.37 (dd, 3J ¼ 8.1 Hz, 4J ¼ 1.6
Hz, 1H, 50-H), 7.44 (dd, 3J ¼ 5.1 Hz, 4J ¼ 1.5 Hz, 1H, 3-Hthienyl),
7.50 (dd, 3J ¼ 8.0 Hz, 4J ¼ 1.0 Hz, 1H, 5-H), 7.52 (dd, 3J ¼ 3.7 Hz, 10 S. Qu and H. Tian, Chem. Commun., 2012, 48, 3039–3051.
4J ¼ 1.5 Hz, 1H, 5-Hthienyl), 7.63 (d, 3J ¼ 8.1 Hz, 1H, 40-H), 7.78 (d, 11 C. B. Nielsen, M. Turbiez and I. McCulloch, Adv. Mater.,
3J ¼ 8.0 Hz, 1H, 4-H), 8.30 (s, 1H, 70-H), 8.31 (s, 1H, 7-H). HR ESI-
2012, 25, 1859–1880.
MS: m/z found for [M + H]+ 623.0859, m/z calculated for [M + H]+ 12 M. J. Robb, S.-Y. Ku, F. G. Brunetti and C. J. Hawker, J. Polym.
623.0846.
Sci., Part A: Polym. Chem., 2013, 51, 1263–1271.
13 C. Guo, B. Sun, J. Quinn, Z. Yan and Y. Li, J. Mater. Chem. C,
2014, 2, 4289–4296.
14 E. D. Głowacki, G. Voss, K. Demirak, M. Havlicek, N. Sunger,
A. C. Okur, U. Monkowius, J. G˛asiorowski, L. Leonat and
N. S. Saricici, Chem. Commun., 2013, 49, 6063–6065.
Conclusions
We have presented synthesis routes to obtain an extended
indigo derivative, 6,60-dithienylindigo (DTI), for application as
an organic semiconductor. We utilize the latent pigment tech- 15 A. Baeyer and V. Drewsen, Chem. Ber., 1882, 15, 2856–2864.
nique of introducing thermolabile solubilizing groups, which 16 P. Susse and R. Wasche, Naturwissenschaen, 1978, 65, 157.
allows chemical manipulation of the otherwise insoluble indigo 17 E. F. Paulus, F. J. J. Leusen and M. U. Schmidt,
moiety. DTI shows promising ambipolar transport behaviour in
CrystEngComm, 2007, 9, 131.
FET devices, with impressive operational stability in air. By 18 Z. Hao and A. Iqbal, Chem. Soc. Rev., 1997, 26, 203.
¨
electropolymerization we could produce poly(DTI). This semi- 19 W. Luttke, H. Hermann and M. Klessinger, Angew. Chem.,
conjugated polymer with indigo in the main chain shows
Int. Ed. Engl., 1966, 5, 598–599.
8096 | J. Mater. Chem. C, 2014, 2, 8089–8097
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry C
¨
20 A. Amat, F. Rosi, C. Miliani, A. Sgamellotti and S. Fantacci, J. 27 E. D. Głowacki, L. Leonat, M. Irimia-Vladu, R. Schwodiauer,
Mol. Struct., 2011, 993, 43–51.
21 D. Jacquemin, J. Preat, V. Wathelet and E. A. Perpete, J.
M. Ullah, H. Sitter, S. Bauer and N. Serdar Saricici, Appl.
Phys. Lett., 2012, 101, 023305.
`
Chem. Phys., 2006, 124, 74104.
22 P. W. Sadler, J. Org. Chem., 1955, 21, 316–318.
28 T. Yamamoto and K. Kizu, J. Phys. Chem., 1995, 99, 8–10.
29 G. Voss and H. Gerlach, Chem. Ber., 1989, 122, 1199–1201.
23 P. W. Sadler and R. L. Warren, J. Am. Chem. Soc., 1956, 78, 30 A. I. Mardare, M. Kaltenbrunner, N. S. Saricici, S. Bauer and
1251–1255. A. W. Hassel, Phys. Status Solidi, 2012, 209, 813–818.
24 H. Zollinger, Color Chemistry. Syntheses, Properties and 31 M. Kraus, S. Richler, A. Opitz, W. Brutting, S. Haas,
¨
Applications of Organic Dyes and Pigments, Wiley-VCH,
Weinheim, 3rd edn, 2003, vol. 43.
T. Hasegawa, A. Hinderhofer and F. Schreiber, J. Appl.
Phys., 2010, 107, 094503.
25 C. M. Cardona, W. Li, A. E. Kaifer, D. Stockdale and 32 G. Voss, Color. Technol., 2006, 122, 317–323.
G. C. Bazan, Adv. Mater., 2011, 23, 2367–2371. 33 A. Suzuki, J. Organomet. Chem., 1999, 576, 147–168.
26 E. D. Głowacki, M. Irimia-Vladu, S. Bauer and N. S. Saricici, 34 J. K. Stille, Angew. Chem., Int. Ed., 1986, 25, 508–524.
J. Mater. Chem. B, 2013, 1, 3742–3753.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. C, 2014, 2, 8089–8097 | 8097